© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25331
    Prev
    Next
    INDEX
    Explanations

    navigation, authentication, and home

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron tends to fire on longer, less frequent “high‐information” tokens (i.e. relatively rare or multi‐syllable words or identifiers).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    );
    -1.14
     dobry
    -1.13
     хорошее
    -1.09
    izzling
    -1.05
    有些人
    -1.05
     ótimo
    -1.04
    \#
    -1.03
     "";
    -1.03
     indivíduos
    -1.02
     [];
    -1.02
    POSITIVE LOGITS
    錚
    1.13
     without
    1.10
    RIO
    1.10
     Colla
    1.05
    赓
    1.03
    Lindo
    0.99
    Epil
    0.98
    APIs
    0.98
    大師
    0.98
     SENSE
    0.98
    Activations Density 0.001%

    No Known Activations